trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 10.0,
  "global_step": 3600,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.39,
      "learning_rate": 1.7222222222222224e-05,
      "loss": 80623473524.736,
      "step": 500
    },
    {
      "epoch": 2.78,
      "learning_rate": 1.4444444444444446e-05,
      "loss": 0.9617,
      "step": 1000
    },
    {
      "epoch": 4.17,
      "learning_rate": 1.1666666666666668e-05,
      "loss": 0.3908,
      "step": 1500
    },
    {
      "epoch": 5.56,
      "learning_rate": 8.888888888888888e-06,
      "loss": 0.3095,
      "step": 2000
    },
    {
      "epoch": 6.94,
      "learning_rate": 6.111111111111112e-06,
      "loss": 0.256,
      "step": 2500
    },
    {
      "epoch": 8.33,
      "learning_rate": 3.3333333333333333e-06,
      "loss": 0.2181,
      "step": 3000
    },
    {
      "epoch": 9.72,
      "learning_rate": 5.555555555555555e-07,
      "loss": 0.1911,
      "step": 3500
    },
    {
      "epoch": 10.0,
      "step": 3600,
      "total_flos": 901255567357440.0,
      "train_loss": 11197704656.541739,
      "train_runtime": 1310.0328,
      "train_samples_per_second": 43.884,
      "train_steps_per_second": 2.748
    }
  ],
  "max_steps": 3600,
  "num_train_epochs": 10,
  "total_flos": 901255567357440.0,
  "trial_name": null,
  "trial_params": null
}